Can't boot image created from panda-x11-base hwpack and linaro-n-ubuntu-desktop on Pandaboard A2

[Ricardo Salveti]

Created an image using the 11.05 release candidate based on panda-x11-base and linaro-n-ubuntu-desktop and I'm unable to boot with my Pandaboard A2.

Same SD card works fine with my EA1 and A1 Pandaboards.

Command used to create the image:
$ sudo linaro-media-create --mmc /dev/sdb --dev panda --rootfs ext3 --hwpack hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz --binary linaro-n-ubuntu-desktop-tar-20110526-0.tar.gz --hwpack-force-yes

Hwpack: http://releases.linaro.org/platform/linaro-n/hwpacks/latest/hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz
Image: http://releases.linaro.org/platform/linaro-n/ubuntu-desktop/latest/linaro-n-ubuntu-desktop-tar-20110526-0.tar.gz

Seems that even X-Loader is not able to initialize, as I can't see any serial nor LED activity.

[Ricardo Salveti (rsalveti)]

Using linaro-image-tools 0.4.7-0ubuntu1~lucid1,

[Alexander Sack (asac)]

is this only a problem of panda-x11-base hwpack? or also the other hwpack? why would the x-loader be different if so?

[Alexander Sack (asac)]

subscribed john as he is the x-loader/u-boot man and this seems to be a bug there (john, please please please unhide your @linaro.org email on launchpad).

Did anyone ever got anthing booting on A2 before?

[warmcat (andy-warmcat)]

There was a 2.3 OMAP4430 CPU version that needed x-loader attention.

There continues to be divergence between x-loader from gitorious and omap-zoom repos.

http://gitorious.org/x-loader/x-loader/commits/master
http://dev.omapzoom.org/?p=bootloader/x-loader.git;a=shortlog;h=refs/heads/omap4_dev

Officially gitorious is meant to be the one place to do x-loader, in fact teams in TI prefer omapzoom. These two patches missing from gitorious look relevant -->

http://dev.omapzoom.org/?p=bootloader/x-loader.git;a=commit;h=ce3a8b47aed6cae62958f32e5be815aec9e171a9
http://dev.omapzoom.org/?p=bootloader/x-loader.git;a=commit;h=1aee45bbcf4f94cd4558315f5a5464ae33ce84b4

[Alexander Sack (asac)]

Thanks. John will not be up for a bit. Do you have an A2 and maybe could help flash this out? Maybe you can provide a MLO build so we can test?

[Alexander Sack (asac)]

(i am at least happy to test that it does not break A1 :))

[warmcat (andy-warmcat)]

I only have Panda A1.

This attached MLO seems to boot my Panda A1 OK.

It's current gitorious x-loader HEAD with the two patches above rebased on it.

[Alexander Sack (asac)]

thanks can you also provide a gitorious HEAD build so we can verify that the issue we see is really in there? otherwise our package might have the bug and we are hunting down the wrong patches ;)

[Alexander Sack (asac)]

i can confirm that it does not regress my A1.

I think our build is 1.5.0+git20110325+b6bbfe7-1ubuntu1 ... e.g. b6bbfe7 ... andy, can you see if your patches apply there too?

[warmcat (andy-warmcat)]

b6bbfe7 is a few patches back from gitorious HEAD so all this will apply cleanly without further changes.

However does it actually fix the A2 issue? Otherwise we are beating on the wrong door.

[Alexander Sack (asac)]

thanks andy!! pfalcon (subscribed) is our tester for A2 now ... stay tuned. hope we have some update in a bit

[Alexander Sack (asac)]

andy, btw, I feel like this are two issues:

 1. DT mess up -> MLO fixes i hope
 2. kernel issues -> we will try ubuntu kernel which ogra claims should work on A2

maybe you can dig around git to see if there are obvious kernel patches missing in linux-linaro for A2?

[Paul Sokolovsky (pfalcon)]

First boot on my Panda A2:

1. linaro-image-tools/linaro-media-create --dev panda --mmc /dev/sdd --binary linaro-natty-nano-tar-20110526-0.tar.gz --hwpack hwpack_linaro-panda_20110526-1_armel_supported.tar.gz
2. Insert card into panda
3. Press button to boot
(No other actions)

Boot log attached.

[Alexander Sack (asac)]

its odd that normal panda hwpack boots :/

On Fri, May 27, 2011 at 12:18 PM, Paul Sokolovsky
<email address hidden> wrote:
> First boot on my Panda A2:
>
> 1. linaro-image-tools/linaro-media-create --dev panda --mmc /dev/sdd --binary linaro-natty-nano-tar-20110526-0.tar.gz --hwpack hwpack_linaro-panda_20110526-1_armel_supported.tar.gz
> 2. Insert card into panda
> 3. Press button to boot
> (No other actions)
>
> Boot log attached.
>
>
> ** Attachment added: "linaro-natty-nano-tar-20110526-0+hwpack_linaro-panda_20110526-1_armel_supported.bootlog.txt"
>   https://bugs.launchpad.net/linaro-images/+bug/788896/+attachment/2144522/+files/linaro-natty-nano-tar-20110526-0%2Bhwpack_linaro-panda_20110526-1_armel_supported.bootlog.txt
>
> --
> You received this bug notification because you are a member of Linaro
> Release Team, which is a direct subscriber.
> https://bugs.launchpad.net/bugs/788896
>
> Title:
>  Can't boot image created from panda-x11-base hwpack and linaro-n
>  -ubuntu-desktop on Pandaboard A2
>
> Status in Linaro Image Builds:
>  New
>
> Bug description:
>  Created an image using the 11.05 release candidate based on
>  panda-x11-base and linaro-n-ubuntu-desktop and I'm unable to boot with
>  my Pandaboard A2.
>
>  Same SD card works fine with my EA1 and A1 Pandaboards.
>
>  Command used to create the image:
>  $ sudo linaro-media-create --mmc /dev/sdb --dev panda --rootfs ext3 --hwpack hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz --binary linaro-n-ubuntu-desktop-tar-20110526-0.tar.gz --hwpack-force-yes
>
>  Hwpack: http://releases.linaro.org/platform/linaro-n/hwpacks/latest/hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz
>  Image: http://releases.linaro.org/platform/linaro-n/ubuntu-desktop/latest/linaro-n-ubuntu-desktop-tar-20110526-0.tar.gz
>
>  Seems that even X-Loader is not able to initialize, as I can't see any
>  serial nor LED activity.
>
> _______________________________________________
> Mailing list: https://launchpad.net/~linaro-release
> Post to     : <email address hidden>
> Unsubscribe : https://launchpad.net/~linaro-release
> More help   : https://help.launchpad.net/ListHelp
>

--

 - Alexander

[warmcat (andy-warmcat)]

Maybe the fix is in x-loader already between b6bbfe7 and HEAD, hwpack may be building HEAD

Good that the kernel is happy though since that means no Ubuntu-specific SAUCE is needed.

Is it certain that power arrangements etc are OK on this A2 that blows chunks?

[Alexander Sack (asac)]

On Fri, May 27, 2011 at 1:33 PM, warmcat <email address hidden> wrote:
> Maybe the fix is in x-loader already between b6bbfe7 and HEAD, hwpack
> may be building HEAD

i hope not ... and checking:

asac@tinya:/tmp/power$ tar tzf
hwpack_linaro-panda_20110526-1_armel_supported.tar.gz | grep x-loa
pkgs/x-loader-omap4-panda_1.5.0+git20110325+b6bbfe7-1ubuntu1_armel.deb
asac@tinya:/tmp/power$ tar tzf
hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz | grep
x-load
pkgs/x-loader-omap4-panda_1.5.0+git20110325+b6bbfe7-1ubuntu1_armel.deb

so negative ... x-loader is identical in both hwpacks :/ ... guess we
need rsalveti to wake up.

also kernel seems to be the same (except that headers is oddly ~ppa3 in x11):

asac@tinya:/tmp/power$ tar tzf
hwpack_linaro-panda_20110526-1_armel_supported.tar.gz | grep linux-
pkgs/linux-image-2.6.38-1003-linaro-omap_2.6.38-1003.4~ppa5_armel.deb
pkgs/linux-image-linaro-omap_2.6.38.1003.4_armel.deb
pkgs/linux-firmware_1.52_all.deb

asac@tinya:/tmp/power$ tar tzf
hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz | grep
linux-
pkgs/linux-headers-2.6.38-1208_2.6.38-1208.11_armel.deb
pkgs/linux-image-2.6.38-1003-linaro-omap_2.6.38-1003.4~ppa5_armel.deb
pkgs/linux-headers-omap4_2.6.38.1208.7_armel.deb
pkgs/linux-libc-dev_2.6.38-8.42_armel.deb
pkgs/linux-headers-2.6.38-1003_2.6.38-1003.4~ppa3_armel.deb
pkgs/linux-firmware_1.52_all.deb
pkgs/linux-image-linaro-omap_2.6.38.1003.4_armel.deb
pkgs/linux-headers-2.6.38-1003-linaro-omap_2.6.38-1003.4~ppa5_armel.deb
pkgs/linux-headers-2.6.38-1208-omap4_2.6.38-1208.11_armel.deb
pkgs/linux-headers-linaro-omap_2.6.38.1003.4_armel.deb
pkgs/linux-headers-2.6.38-8_2.6.38-8.42_all.deb

--

 - Alexander

[warmcat (andy-warmcat)]

Well since bug report is nothing from X-loader even, kernel needn't be worried about.

Maybe his SD card slot is broken, or power arrangements issue... have to hear if he can downgrade to anything workable using that card (or has ever seen this Panda A2 work in fact).

[Alexander Sack (asac)]

On Fri, May 27, 2011 at 2:02 PM, warmcat <email address hidden> wrote:
> Well since bug report is nothing from X-loader even, kernel needn't be
> worried about.
>
> Maybe his SD card slot is broken, or power arrangements issue... have to
> hear if he can downgrade to anything workable using that card (or has
> ever seen this Panda A2 work in fact).

you are right. i am hoping for pfalcon finding a 4G sdcard ... or
someone else who has A2

--

 - Alexander

[Ricardo Salveti (rsalveti)]

A2 is still ES2.1, so our X-loader and Kernel should be fine.

When testing yesterday with the nano image it worked OK, I'm now trying it again.

[Paul Sokolovsky (pfalcon)]

Next test:

1. linaro-image-tools/linaro-media-create --dev panda --mmc /dev/sdd --binary linaro-n-ubuntu-desktop-tar-20110526-0.tar.gz --hwpack hwpack_linaro-panda-x11-base_20110526-0_armel_supported.tar.gz
2. Boot from card
3. See serial boot ok
4. Plug HDMI
5. See blank screen (but monitor doesn't power off)
6. Reboot with HDMI plugged
7. See X start up (a bit long have to say ;-) )
8. Attach USB mouse, keyboard - both work

[Ricardo Salveti (rsalveti)]

Paul, is this test result with an A2?

[Paul Sokolovsky (pfalcon)]

Afterwards:

echo "mem">/sys/power/state screen blanks (but monitor doesn't power off), heartbeat led still blinks, keypress, mouse click, GPIO_121 button don't wake up (nothing wakes up).

[Paul Sokolovsky (pfalcon)]

After more idle time, then visiting in Firefox multimedia site which apparently would have lotsa flash:

root@linaro:~# [ 2745.864013] omapdss DISPC error: GFX_FIFO_UNDERFLOW, disabling GFX
[ 2745.930267] omapdss DISPC error: SYNC_LOST_DIGIT, disabling TV
[ 2746.033569] omapdss DISPC error: timeout waiting for EVSYNC
[ 2746.135162] omapdss DISPC error: timeout waiting for EVSYNC
[ 2746.400848] omapdss DISPC error: timeout waiting for EVSYNC
[ 2746.502807] omapdss DISPC error: timeout waiting for EVSYNC

HDMI output goes blank here.

[Paul Sokolovsky (pfalcon)]

Ok (this goes offtopic on this bug, sorry), the GFX shutdown above is 100% reproducible by visiting with in-image firefox the site http://101.ru (just random flashy-spammy internet radio portal).

[Ricardo Salveti (rsalveti)]

Paul, this may be bug 746133.

Now about this bug, flashed 4 different SD cards with hwpack_linaro-panda_20110526-1_armel_supported.tar.gz and linaro-natty-nano-tar-20110526-0.tar.gz (kingston class 4, micro center class 6, patriot lx class 10, patriot class4) and only 2 worked with my A2 (micro center and patriot lx). Interesting enough that all 4 cards worked fine with A1 and EA1, so it could be a hardware problem.

[John Rigby (jcrigby)]

No idea if this is relevant but in the batch of panda's that mwaddel had at UDS a number had bad sd slots. The spring loaded contacts were not making good contact with the card. But I have no idea why this would depend on the image.

[Ricardo Salveti (rsalveti)]

This could be the reason, as I was able to reproduce the issue with different images (nano, developer and desktop) with some of my sd cards, but not all.